Scanner comparison test

  • Thread starter Thread starter Jim Hutchison
  • Start date Start date
This could really be a great resource considering everyone will basically be
using the same test target. I hope you get lots of participants with a wide
variety of scanner types.

Doug
 
more than 40 participants so far...

Woo!



This could really be a great resource considering everyone will basically be
using the same test target. I hope you get lots of participants with a wide
variety of scanner types.

Doug
Click to expand...

jim h


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

http://www.jamesphotography.ca

-free downloads
-prize for every 1,000th visitor
-scanning service

Even a bit of humour...
 
Jim said:
more than 40 participants so far...
Click to expand...

A thought, since signing up - will it be necessary to send back a 50+Mb
TIFF with the results? Or can it be jpegised? The former isn't very
practical over a 33.6kbps connection!
 
You'll get instructions with the slide. But ya, please save it as a
jpeg at the highest quality.

With only a 33.6, you may want to consider burning it to CD, but I'll
leave that up to you...


jim


A thought, since signing up - will it be necessary to send back a 50+Mb
TIFF with the results? Or can it be jpegised? The former isn't very
practical over a 33.6kbps connection!
Click to expand...

jim h


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

http://www.jamesphotography.ca

-free downloads
-prize for every 1,000th visitor
-scanning service

Even a bit of humour...
 
Jim Hutchison said:
Goodness no - I've mailed to the Netherlands and Australia so far...
Click to expand...

Would be great if one went out to Germany too ;-)

regards
Markus
 
I suppose USA only? I'm in Europe...
Goodness no - I've mailed to the Netherlands and Australia so far...
Click to expand...

Excellent! I was wondering about that, too (Italy).
Any timeframe available for the shipping? (I can't wait to start the test) :)

Thanks!

Fernando
 
You should have it real soon...


Excellent! I was wondering about that, too (Italy).
Any timeframe available for the shipping? (I can't wait to start the test) :)

Thanks!

Fernando
Click to expand...

jim h


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

http://www.jamesphotography.ca

-free downloads
-prize for every 1,000th visitor
-scanning service

Even a bit of humour...
 
Update: Over 70 applications; all have been mailed a slide. I've
received over 20 results back so far.

There's allot of interest in this! So far, it's quite obvious from
the results that you do indeed get what you pay for! Cheap scanner =
cheap results.

I have a "News" page on my web site for updates such as this...




Anyone interested in a scanner bake-off, see
http://www.jamesphotography.ca/scanner_test.html for participation
specifics.

Cya!
Click to expand...

jim h


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

http://www.jamesphotography.ca

-free downloads
-prize for every 1,000th visitor
-scanning service

Even a bit of humour...
 
Update: Over 70 applications; all have been mailed a slide. I've
received over 20 results back so far.

There's allot of interest in this! So far, it's quite obvious from
the results that you do indeed get what you pay for! Cheap scanner =
cheap results.

I have a "News" page on my web site for updates such as this...
Click to expand...

Any preliminary findings on the Minolta Dimage Scan Dual IV?

I've weighing my options on this model verses a Fuji S7000.

Thanks!
Darrell
 
Jim Hutchison said:
Update: Over 70 applications; all have been mailed a slide. I've
received over 20 results back so far.
Click to expand...

So far, MTF results don't seem to be a good predictor of visual appearance.
Thus, here are the more-respected scanners sorted by chromatic aberration:

Nikon 8000 4000 manual Silverfast 26.92 .26 2.84
Canon FS4000 4000 auto Filmget 22.2 .373 3.03
Minolta IV 3200 auto Vuescan 20.5 .386 2.8
Nikon LS-40 2900 auto ?????? 17.12 .403 2.8
Polaroid 4000 4000 auto Silverfast 23.46 .426 3.52
Minolta DSMP 4800 manual Vuescan 21.9 .468 4.36
Nikon LS-40 2900 manual Vuescan 18 .471 2.67
Nikon 4000 4000 manual Nikonscan 22.36 .495 3.38
Minolta IV 3200 auto Minolta 20.91 .54 2.27
Minolta IV 3200 auto Minolta 17.9 .572 2.47
Nikon 8000 4000 ??? Nikonscan3.1 18.7 .592 4.29
Nikon LS-50 4000 auto Nikonscan4 26.29 .697 2.88
Minolta 5400 5400 auto Vuescan8 21.3 .754 4.97
Nikon 4000 4000 auto Vuescan7.6 20.6 .833 5.04
Canon FS4000 4000 auto Filmget1.04 20.3 .899 3.25
Canon FS4000 4000 ??? Vuescan8.0.1 19.84 .947 3.62
Minolta 5400 5400 manual Vuescan 23.8 .975 4.04
Nikon 5000 4000 auto Nikonscan 22.4 .98 3.00
Minolta DSMP 4800 manual Minolta1.1.3 19.65 .987 4.21
Canon FS4000 4000 auto Filmget1.0 21.9 1.45 3.64

Interesting that the Canon FS-4000 is in both 2nd place and last place!
Dirty optics, perhaps. Also interesting how well the Minolta Dimage IV
performs relative to more expensive scanners.

Now here are the more-respected scanners sorted by 10-90% pixel rise
(hmm, there's that pesky Minolta Dimage IV again):

Minolta IV 3200 auto Minolta 20.91 .54 2.27
Minolta IV 3200 auto Minolta 17.9 .572 2.47
Nikon LS40 2900 manual Vuescan 18 .471 2.67
Minolta IV 3200 auto Vuescan 20.5 .386 2.8
Nikon LS40 2900 auto ???????? 17.12 .403 2.8
Nikon 8000 4000 manual Silverfast 26.92 .26 2.84
Nikon LS50 4000 auto Nikonscan4 26.29 .697 2.88
Nikon 5000 4000 auto Nikonscan 22.4 .98 3.00
Canon FS4000 4000 auto Filmget 22.2 .373 3.03
Canon FS4000 4000 auto Filmget1.04 20.3 .899 3.25
Nikon 4000 4000 manual Nikonscan 22.36 .495 3.38
Polaroid 4000 4000 auto Silverfast 23.46 .426 3.52
Canon FS4000 4000 ??? Vuescan8.0.1 19.84 .947 3.62
Canon FS4000 4000 auto Filmget1.0 21.9 1.45 3.64
Minolta 5400 5400 manual Vuescan 23.8 .975 4.04
Minolta DSMP 4800 manual Minolta1.1.3 19.65 .987 4.21
Nikon 8000 4000 ??? Nikonscan3.1 18.7 .592 4.29
Minolta DSMP 4800 manual Vuescan 21.9 .468 4.36
Minolta 5400 5400 auto Vuescan8 21.3 .754 4.97
Nikon 4000 4000 auto Vuescan7.6 20.6 .833 5.04
 
Bill Tuthill said:
So far, MTF results don't seem to be a good predictor of visual
appearance.
Thus, here are the more-respected scanners sorted by chromatic
aberration:
Click to expand...
Interesting that the Canon FS-4000 is in both 2nd place and last
place!
Dirty optics, perhaps. Also interesting how well the Minolta Dimage
IV
performs relative to more expensive scanners.

Now here are the more-respected scanners sorted by 10-90% pixel rise
(hmm, there's that pesky Minolta Dimage IV again):
Click to expand...
<snip>

Your sorting order may be interesting, but I see at least one mistake:

Minolta IV 3200 auto Vuescan 20.5 .386 2.8

was probably the one representing my results. I didn't use Vuescan
however!
 
So far, MTF results don't seem to be a good predictor of visual appearance.
Thus, here are the more-respected scanners sorted by chromatic aberration:

Now here are the more-respected scanners sorted by 10-90% pixel rise
(hmm, there's that pesky Minolta Dimage IV again):
Click to expand...

You may be right if you like looking at scans at 100% on your screen. If you
want to print, you have to take resolution into account.

For example, my LS-4000 gets CA .495 and rise 3.38, and my LS-2000 gets
CA: .43 and rise 2.65. However, the LS-4000 is quite a bit better than
the LS-2000.
 
Bill Tuthill said:
So far, MTF results don't seem to be a good predictor of visual
Click to expand...
appearance.

MTF measures sensor response (per pixel) in the frequency domain.
A single number, such as MTF at 50% modulation, cannot tell the whole story,
especially since we know that an image on film after scanning has a very
different MTF shape (lower modulation at medium frequencies, higher at high
frequencies) than a digicam (higher modulation at medium frequencies, lower
at, and limited by, the lower Nyquist frequency).
IMHO an MTF20 is more relevant for scanned film.

SNIP
Interesting that the Canon FS-4000 is in both 2nd place and last place!
Click to expand...

The small size of the slanted edge, and graininess/noise can cause larger
variations, even with the same scan. Differences in gamma also change the
result.
Now here are the more-respected scanners sorted by 10-90% pixel rise
Click to expand...
SNIP

The 10-90% rise (just like chromatic aberration) doesn't weigh in the number
of pixels or sampling density. For each pixel of e.g. a 3200ppi scanner, a
5400ppi scanner can use 1.7 pixels, so with equal optics, edge transitions
wiil be wider with a higher ppi scanner.

Since sensors differ in size and projection magnification, a better metric
would be the LW/PH (line widths per picture height). However, for that,
cropping would have to be exactly equal, or image height can be manually
pre-set to 24mm / 25.4 * ppi .

Bart
 
Philip Homburg said:
You may be right if you like looking at scans at 100% on your screen.
If you want to print, you have to take resolution into account.

For example, my LS-4000 gets CA .495 and rise 3.38, and my LS-2000 gets
CA: .43 and rise 2.65. However, the LS-4000 is quite a bit better than
the LS-2000.
Click to expand...

Right. I was mostly trolling, though to my eye the non-black blacks
produced by the Minolta 5400 are downright disturbing. This bake-off
shows me that the 5400 isn't all that great.

Hopefully we'll soon see a larger sample size for the Coolscan V (LS-50)
which seems to be the best value (with ICE) for the money right now.
 
Back
Top